- Mechanical and electrical responsible for improvements on industrial machinery.- Responsible for digital improvements on the company (2 people in charge)

- Underwater simulation using Gazebo simulator for the UUV based on Dave Project. It includes necessary plugins, such as hydrostatic forces and thrusters.

- Project leader of the UUV (17 developers): Mechanics, Electronics, Perception, Localization and Control.- Actual maintainer of the UUV's ROS repository.

- Underwater deep learning computer vision using the underwater image formation model and synthetic data from diffuse models.- Evaluation of 3d Gaussian Splats and Neural Radiance Fields for underwater computer vision tasks.

- Heatmaps implementation for deep neural networks to explain the focus point during the detection task.- Constant improvement of the synthetic data generation strategy based on Heatmaps, feature maps and metrics analysis.- Pipeline implementation for synthetic data generation using deep diffuse models based on Low-Rank Adaptation of Large Language Models (LoRAs).

Title: Object Detection System Implementation for an Autonomous Mobile Robot using a Synthetic 2D-RGB Data Generation Approach- Exploring the domain randomization technique to optimize the performance of neural networks,based on approaches written in several computer vision papers.

Object detection using convolutional neural networks (CNNs) for an industrial environment. Analysisof different CNNs and hyperparameter optimization.

- Usage of CNNs for object detection and metrics analysis.- Robotic Operating System (ROS) integration of distinct packages of the UUV project and tests on Gazebo.- Third place overall on the Robosub 2020 competition.